1912] Ritter: The Marine Biological Station of San Diego 181 
locations be selected as far as possible, which can be fixed by 
accurate sextant observations on points of land. Second, the 
plan must be made with reference to some particular correlation 
problem or small group of closely interdependent problems, for 
it is obvious that all such problems cannot be investigated at the 
same time. The problems fall into a natural sequence so that 
those which are selected for investigation at a given time will 
depend upon work already accomplished. 
Suppose we choose variation with respect to season as our 
problem. Obviously to compare the data obtained from one 
locality during the summer with those obtained from another 
locality during the spring, fall or winter, would come short of 
the best scientific procedure. Were comparison thus made, the 
variations observed in the kinds and abundance of organisms 
might be due to differences in locality and not to the effects of 
season. The first point of importance then, is to confine investi- 
vation to the same localities during all the seasons. Again, data 
obtained during daylight in summer would not reveal seasonal 
effect with certainty if compared with those obtained during 
twilight or darkness in the other seasons, because the variations 
noted might be due entirely to variations in light. Consequently 
collecting at one season must be done under approximately the 
same daylight conditions each day for all the other seasons. 
Similar remarks apply to problems of variation with tempera- 
ture, currents, density, gas-content, and other hydrographic 
factors, as well as to a number of meteorological factors such as 
clouds, fog, rain, ete. However an important difference exists, 
namely, that while we can confine collecting to definite localities 
and times of day we cannot control or even foresee, except in a 
general way, what the hydrographic and meteorological condi- 
tions will be. Nevertheless it is necessary to eliminate the vitiat- 
ing effect of the numerous variable factors if we are to determine 
the effeets of season. This can be accomplished only by repeti- 
tion of the observations under the same conditions, so far as the 
means of observing are concerned. To be adequate then, our 
plan must provide for frequently repeated collecting with the 
same apparatus, in the same depths, at the same localities, and 
at the corresponding time of day during all the seasons. 
